WEBVTT

NOTE
Transcription provided by Deepgram

00:00:01.280 --> 00:00:04.319
<v Speaker 0>Hello, and welcome to the Geek News Central

00:00:04.319 --> 00:00:09.360
<v Speaker 0>podcast. It is Sunday, October twenty sixth. This

00:00:09.360 --> 00:00:12.400
<v Speaker 0>is episode eighteen forty nine. We got a

00:00:12.400 --> 00:00:14.000
<v Speaker 0>pretty sweet episode lined up for you guys

00:00:14.000 --> 00:00:17.455
<v Speaker 0>tonight. This is Ray speaking this, this afternoon,

00:00:18.555 --> 00:00:20.815
<v Speaker 0>afternoon for me at least, doing a little

00:00:21.035 --> 00:00:23.275
<v Speaker 0>free shift show. So, hopefully, I can get

00:00:23.275 --> 00:00:25.275
<v Speaker 0>this nice and polished before I have to

00:00:25.275 --> 00:00:27.675
<v Speaker 0>head off to my, one of my jobs.

00:00:27.675 --> 00:00:32.360
<v Speaker 0>But, lead story for you folks this evening.

00:00:33.380 --> 00:00:36.820
<v Speaker 0>Eight hundred plus tech leaders urge pause on

00:00:36.820 --> 00:00:40.980
<v Speaker 0>superintelligence race. We've seen the back and forth

00:00:40.980 --> 00:00:44.375
<v Speaker 0>on this over and over. We've seen the

00:00:44.375 --> 00:00:49.035
<v Speaker 0>war between AI doomers and AI boomers. But,

00:00:49.735 --> 00:00:51.574
<v Speaker 0>a little bit of context into this article

00:00:51.574 --> 00:00:55.110
<v Speaker 0>from CNBC before we get into it. Over

00:00:55.110 --> 00:00:58.390
<v Speaker 0>eight hundred notable figures, including Apple cofounder Steve

00:00:58.390 --> 00:01:03.030
<v Speaker 0>Wozniak, Virgin Mobile or Virgin's Richard Branson, and

00:01:03.030 --> 00:01:05.590
<v Speaker 0>even members of the UK royal family have

00:01:05.590 --> 00:01:08.645
<v Speaker 0>signed a petition petition urging world governments to

00:01:08.645 --> 00:01:11.704
<v Speaker 0>slow or pause the development of superintelligent AI.

00:01:12.325 --> 00:01:15.045
<v Speaker 0>The call follows recent breakthroughs in autonomous reasoning

00:01:15.045 --> 00:01:17.545
<v Speaker 0>and fears that AI may surpass human control

00:01:17.604 --> 00:01:22.110
<v Speaker 0>faster than regulation can keep up. Kinda interesting.

00:01:22.110 --> 00:01:23.310
<v Speaker 0>I mean, I feel like this marks one

00:01:23.310 --> 00:01:26.030
<v Speaker 0>of the most unified public statements from Tech

00:01:26.030 --> 00:01:30.350
<v Speaker 0>Pioneers warning of AI's rapid trajectory. Whether you're,

00:01:30.350 --> 00:01:33.230
<v Speaker 0>like, an optimist or skeptic regarding AI, this

00:01:33.230 --> 00:01:36.174
<v Speaker 0>kinda shows that even insiders are nervous about

00:01:36.174 --> 00:01:38.975
<v Speaker 0>what's being, like, unleashed, and and that a

00:01:38.975 --> 00:01:41.615
<v Speaker 0>pause might be less about fear and more

00:01:41.615 --> 00:01:45.055
<v Speaker 0>about keeping accountability for everything and how rapidly

00:01:45.055 --> 00:01:49.299
<v Speaker 0>it's growing. But some of the key points

00:01:49.299 --> 00:01:51.700
<v Speaker 0>from this article, superintelligence, a hypothetical form of

00:01:51.700 --> 00:01:53.860
<v Speaker 0>AI that surpasses human intelligence, has become a

00:01:53.860 --> 00:01:56.420
<v Speaker 0>buzzword in the AI race between giants like

00:01:56.420 --> 00:02:03.165
<v Speaker 0>Meta and OpenAI. Yep. A group of prominent

00:02:03.165 --> 00:02:06.045
<v Speaker 0>figures, including artificial intelligence and technology experts, has

00:02:06.045 --> 00:02:07.805
<v Speaker 0>called for an end to efforts up to

00:02:07.805 --> 00:02:10.365
<v Speaker 0>create superintelligence, a form of AI that would

00:02:10.365 --> 00:02:14.519
<v Speaker 0>surpass humans on essentially all cognitive tasks. Over

00:02:14.519 --> 00:02:17.239
<v Speaker 0>eight hundred fifty people, including tech leaders like

00:02:17.239 --> 00:02:20.859
<v Speaker 0>Virgin Group founder Richard Branson and Apple, cofounder

00:02:21.079 --> 00:02:24.200
<v Speaker 0>Steve Wozniak signed a statement published Wednesday calling

00:02:24.200 --> 00:02:26.219
<v Speaker 0>for a pause in the development of superintelligence.

00:02:28.465 --> 00:02:30.625
<v Speaker 0>The list of signatories was notably topped by

00:02:30.625 --> 00:02:34.245
<v Speaker 0>prominent AI pioneers, including this computer scientist Yoshua

00:02:34.305 --> 00:02:37.925
<v Speaker 0>Bengio and Joffe Hinton, who are widely considered

00:02:37.985 --> 00:02:42.320
<v Speaker 0>godfathers godfathers of modern AI, meaning AI researchers

00:02:42.540 --> 00:02:46.160
<v Speaker 0>like UC Berkeley Stewart Russell also signed on.

00:02:49.900 --> 00:02:52.560
<v Speaker 0>That's kinda kinda interesting to kinda see this

00:02:52.620 --> 00:02:57.015
<v Speaker 0>this front forming, against AI, here coming into

00:02:57.015 --> 00:02:58.715
<v Speaker 0>the end of the year. You know, we've

00:02:58.855 --> 00:03:02.055
<v Speaker 0>seen a lot of push for it, leading

00:03:02.055 --> 00:03:04.215
<v Speaker 0>up to this, and I feel like the

00:03:04.215 --> 00:03:07.735
<v Speaker 0>pushback has been real quiet, mostly amongst people

00:03:07.735 --> 00:03:13.680
<v Speaker 0>that recognize the, the implications societally that building,

00:03:14.380 --> 00:03:16.860
<v Speaker 0>AI supercenters and whatnot are causing in a

00:03:16.860 --> 00:03:22.220
<v Speaker 0>lot of our communities. So it's cool to

00:03:22.220 --> 00:03:26.845
<v Speaker 0>see them trying to pump slowdowns on, a

00:03:26.845 --> 00:03:31.265
<v Speaker 0>lot of this superintelligence, especially being that, like,

00:03:31.565 --> 00:03:36.625
<v Speaker 0>our infrastructures are not built to support these

00:03:38.980 --> 00:03:42.660
<v Speaker 0>the supercenters for AI. I mean, they're reporting

00:03:42.660 --> 00:03:46.180
<v Speaker 0>these centers in, like, gigawatts now instead of,

00:03:46.180 --> 00:03:48.980
<v Speaker 0>like, their cost to, like, almost, like, signal

00:03:48.980 --> 00:03:51.734
<v Speaker 0>that that that we need more energy on

00:03:51.734 --> 00:03:54.075
<v Speaker 0>the grid. This is not most of these

00:03:54.855 --> 00:03:56.454
<v Speaker 0>plans that they have in place, the the

00:03:56.534 --> 00:03:59.575
<v Speaker 0>with energy that we're putting on the grid

00:03:59.575 --> 00:04:01.734
<v Speaker 0>is not enough to support those supercenters. So

00:04:01.734 --> 00:04:03.990
<v Speaker 0>kinda interesting to see what's gonna change with

00:04:03.990 --> 00:04:08.709
<v Speaker 0>that, coming this year, but pretty sweet, news

00:04:08.709 --> 00:04:11.370
<v Speaker 0>from CNBC. This one's written by Dylan Butts.

00:04:13.670 --> 00:04:17.235
<v Speaker 0>But, yeah, let's get into the show. A

00:04:17.235 --> 00:04:19.155
<v Speaker 0>little bit of new stuff with me this

00:04:19.155 --> 00:04:21.555
<v Speaker 0>week. Had a real real busy school week,

00:04:21.555 --> 00:04:25.475
<v Speaker 0>and, we finally were able to get a

00:04:25.475 --> 00:04:27.475
<v Speaker 0>little beta running for some of the changes

00:04:27.475 --> 00:04:29.555
<v Speaker 0>at Blueberry. So I'm super excited to kinda

00:04:29.555 --> 00:04:31.725
<v Speaker 0>see that going forward. Hopefully, the test rollout,

00:04:31.725 --> 00:04:33.590
<v Speaker 0>and you guys can see the PowerPress changes

00:04:33.590 --> 00:04:35.990
<v Speaker 0>for those of you that podcast and use

00:04:35.990 --> 00:04:40.630
<v Speaker 0>PowerPress. But, outside of that, just I've had

00:04:40.630 --> 00:04:42.650
<v Speaker 0>a super busy week, a lot of homework,

00:04:44.805 --> 00:04:47.045
<v Speaker 0>just juggling the the two jobs right now.

00:04:47.045 --> 00:04:48.725
<v Speaker 0>So just trying to keep my head on

00:04:48.725 --> 00:04:56.165
<v Speaker 0>straight. But, less less about me. Let's talk

00:04:56.165 --> 00:04:59.030
<v Speaker 0>about, our amazing sponsor. Big shout out to

00:04:59.030 --> 00:05:02.870
<v Speaker 0>our incredible sponsor, GoDaddy. Score exclusive deals and

00:05:02.870 --> 00:05:04.710
<v Speaker 0>discounts right now at geek news central dot

00:05:04.710 --> 00:05:07.510
<v Speaker 0>com forward slash godaddy. And to all of

00:05:07.510 --> 00:05:09.430
<v Speaker 0>our listeners, you are truly the heartbeat of

00:05:09.430 --> 00:05:12.255
<v Speaker 0>this show. If you wanna support Geek News

00:05:12.255 --> 00:05:14.255
<v Speaker 0>Central directly, head over to geek news central

00:05:14.255 --> 00:05:16.495
<v Speaker 0>dot com forward slash insider and become a

00:05:16.495 --> 00:05:19.535
<v Speaker 0>GNC insider. And then on top of that,

00:05:19.535 --> 00:05:21.455
<v Speaker 0>you're gonna wanna grab yourself a modern podcast

00:05:21.455 --> 00:05:23.535
<v Speaker 0>app at podcast apps dot com for the

00:05:23.535 --> 00:05:26.910
<v Speaker 0>full next gen podcast experience. Of course, you

00:05:26.910 --> 00:05:28.750
<v Speaker 0>can check out the entire Geek News Central

00:05:28.750 --> 00:05:31.230
<v Speaker 0>universe at geek news central dot com, where

00:05:31.230 --> 00:05:33.550
<v Speaker 0>you'll find all our partner shows and tons

00:05:33.550 --> 00:05:35.950
<v Speaker 0>of great content published every single week. We

00:05:35.950 --> 00:05:37.560
<v Speaker 0>have slowed down a little bit on the,

00:05:39.185 --> 00:05:42.385
<v Speaker 0>the posts in general. I am kinda seeing

00:05:42.385 --> 00:05:43.665
<v Speaker 0>if we can get some of our writers

00:05:43.665 --> 00:05:45.825
<v Speaker 0>back to bring, some of the posts back,

00:05:45.825 --> 00:05:47.825
<v Speaker 0>but we're we're still kind of in an

00:05:47.825 --> 00:05:50.565
<v Speaker 0>in between place in terms of the business

00:05:50.625 --> 00:05:54.270
<v Speaker 0>and the show. And there's a whole process

00:05:54.270 --> 00:05:56.030
<v Speaker 0>in Michigan that you have to follow, like

00:05:56.030 --> 00:05:57.950
<v Speaker 0>a probate process that makes it a little

00:05:57.950 --> 00:06:02.590
<v Speaker 0>bit more difficult to actually reallocate or use

00:06:02.590 --> 00:06:06.885
<v Speaker 0>those funds. So for the time being, the

00:06:06.885 --> 00:06:08.885
<v Speaker 0>business funds are kinda kinda frozen, and we

00:06:08.885 --> 00:06:10.325
<v Speaker 0>just have to wait it out until the

00:06:10.325 --> 00:06:16.005
<v Speaker 0>probate's over. So look forward to more articles

00:06:16.005 --> 00:06:18.005
<v Speaker 0>in the future once we kinda get everything

00:06:18.005 --> 00:06:25.080
<v Speaker 0>settled on my father's side. But, yeah, stay

00:06:25.080 --> 00:06:26.760
<v Speaker 0>in the loop. Follow or subscribe in your

00:06:26.760 --> 00:06:28.520
<v Speaker 0>favorite podcast app, and don't forget to sign

00:06:28.520 --> 00:06:30.280
<v Speaker 0>up for our newsletter. I do try to

00:06:30.280 --> 00:06:32.840
<v Speaker 0>send the newsletters out after every show. I

00:06:32.840 --> 00:06:35.800
<v Speaker 0>still have to get Chris, access to the

00:06:35.800 --> 00:06:37.935
<v Speaker 0>Geek News email so that he can kinda

00:06:37.935 --> 00:06:41.295
<v Speaker 0>send those out too. But we are working

00:06:41.295 --> 00:06:44.895
<v Speaker 0>on getting that consistently done. So, hopefully, if

00:06:44.895 --> 00:06:46.655
<v Speaker 0>you sign up for the newsletter, you're seeing

00:06:46.655 --> 00:06:50.095
<v Speaker 0>those coming out and, respond. If you want

00:06:50.095 --> 00:06:51.600
<v Speaker 0>to, tell tell me what you think. If

00:06:51.600 --> 00:06:52.960
<v Speaker 0>you want me to change up the format,

00:06:52.960 --> 00:06:54.660
<v Speaker 0>if you want me to do more articles,

00:06:54.880 --> 00:06:57.840
<v Speaker 0>please let me know. We really, really wanna

00:06:57.840 --> 00:07:00.160
<v Speaker 0>hear your feedback. You can always reach us

00:07:00.160 --> 00:07:02.400
<v Speaker 0>directly at geek news at g mail dot

00:07:02.400 --> 00:07:05.305
<v Speaker 0>com. Now we wanna take a moment to

00:07:05.305 --> 00:07:07.865
<v Speaker 0>thank GoDaddy, our longtime sponsor for their incredible

00:07:07.865 --> 00:07:10.025
<v Speaker 0>support of this podcast. They make it possible

00:07:10.025 --> 00:07:11.545
<v Speaker 0>for us to keep doing what we love

00:07:11.545 --> 00:07:14.665
<v Speaker 0>and what you enjoy hearing each week. If

00:07:14.665 --> 00:07:16.265
<v Speaker 0>you're ready to start your own website or

00:07:16.265 --> 00:07:18.845
<v Speaker 0>even launch a podcast, GoDaddy has you covered.

00:07:19.250 --> 00:07:21.250
<v Speaker 0>At geek news central dot com forward slash

00:07:21.250 --> 00:07:23.570
<v Speaker 0>GoDaddy, you'll find special promo codes that make

00:07:23.570 --> 00:07:27.990
<v Speaker 0>it easy and affordable to get started. Economy

00:07:28.050 --> 00:07:30.290
<v Speaker 0>hosting, just six ninety nine a month. That's

00:07:30.290 --> 00:07:31.890
<v Speaker 0>a full year of hosting with a free

00:07:31.890 --> 00:07:35.065
<v Speaker 0>domain name, professional email, and an SSL certificate.

00:07:35.845 --> 00:07:37.525
<v Speaker 0>And if you're like us here at Geek

00:07:37.525 --> 00:07:40.645
<v Speaker 0>News Central and you love WordPress, WordPress hosting,

00:07:40.645 --> 00:07:42.885
<v Speaker 0>only twelve ninety nine a month, including that

00:07:42.885 --> 00:07:46.265
<v Speaker 0>same free domain name, email, and SSL certificate,

00:07:46.819 --> 00:07:49.300
<v Speaker 0>Get started podcasting today. You can post your

00:07:49.300 --> 00:07:54.900
<v Speaker 0>own show right on the website. And then

00:07:54.900 --> 00:07:56.180
<v Speaker 0>if you just need a domain name, you

00:07:56.180 --> 00:07:57.539
<v Speaker 0>wanna grab one ahead of time, maybe you're

00:07:57.539 --> 00:07:59.699
<v Speaker 0>thinking about starting a business, starting something up,

00:07:59.699 --> 00:08:01.574
<v Speaker 0>you got family that's in you know, that's

00:08:01.574 --> 00:08:03.895
<v Speaker 0>building out portfolios, they like to do artistic

00:08:03.895 --> 00:08:05.655
<v Speaker 0>stuff, they wanna share with the world, get

00:08:05.655 --> 00:08:07.655
<v Speaker 0>a domain name today for eleven ninety nine.

00:08:07.655 --> 00:08:10.455
<v Speaker 0>It's a really great way to lock that

00:08:10.455 --> 00:08:11.735
<v Speaker 0>in so that in the future, if you

00:08:11.735 --> 00:08:13.815
<v Speaker 0>wanna build out your website, you can do

00:08:13.815 --> 00:08:15.800
<v Speaker 0>that really easily. And even if you're not

00:08:15.800 --> 00:08:18.919
<v Speaker 0>too super tech savvy, GoDaddy's website builder gives

00:08:18.919 --> 00:08:21.160
<v Speaker 0>you a free thirty day trial on personal

00:08:21.160 --> 00:08:22.919
<v Speaker 0>business and business plus plans, so you can

00:08:22.919 --> 00:08:26.840
<v Speaker 0>get online in no time. We use GoDaddy

00:08:26.840 --> 00:08:29.560
<v Speaker 0>products here at Geek News Central. I personally

00:08:29.560 --> 00:08:32.285
<v Speaker 0>use, GoDaddy products for my own website, and

00:08:32.285 --> 00:08:34.285
<v Speaker 0>I can tell you that their reliability is

00:08:34.285 --> 00:08:37.565
<v Speaker 0>the real deal. Your support at GoDaddy directly

00:08:37.565 --> 00:08:40.385
<v Speaker 0>supports this show, so please use our codes,

00:08:40.685 --> 00:08:43.505
<v Speaker 0>click and save, and help this independent podcast

00:08:43.645 --> 00:08:46.180
<v Speaker 0>going strong. Share those links with friends and

00:08:46.180 --> 00:08:48.180
<v Speaker 0>family. Every time one's used, it's like writing

00:08:48.180 --> 00:08:50.500
<v Speaker 0>a check to the show. We truly thank

00:08:50.500 --> 00:08:52.260
<v Speaker 0>you for supporting us, and a huge thanks

00:08:52.260 --> 00:08:54.500
<v Speaker 0>to GoDaddy for believing in independent creators like

00:08:54.500 --> 00:08:59.445
<v Speaker 0>us here at Geek News Central. Now I'm

00:08:59.445 --> 00:09:02.185
<v Speaker 0>sure you guys have all heard about the

00:09:02.325 --> 00:09:06.265
<v Speaker 0>crazy AWS outage we had earlier this week.

00:09:08.005 --> 00:09:11.360
<v Speaker 0>Luckily, we weren't we didn't seem too affected

00:09:11.360 --> 00:09:13.620
<v Speaker 0>at Blueberry. I had some issues with Zoom,

00:09:14.000 --> 00:09:15.840
<v Speaker 0>that morning, just trying to, like, get into

00:09:15.840 --> 00:09:18.960
<v Speaker 0>meetings to load stuff up. But, outside of

00:09:18.960 --> 00:09:20.880
<v Speaker 0>it being slow, it it wasn't that wasn't

00:09:20.880 --> 00:09:25.654
<v Speaker 0>a huge issue. But I I found this

00:09:25.654 --> 00:09:28.214
<v Speaker 0>article that says, you know, Amazon reportedly placed

00:09:28.214 --> 00:09:31.735
<v Speaker 0>replaced forty percent of AWS DevOps workers with

00:09:31.735 --> 00:09:34.555
<v Speaker 0>AI days before the major crash, which is,

00:09:35.975 --> 00:09:37.334
<v Speaker 0>I mean, I feel like that would be

00:09:37.334 --> 00:09:42.900
<v Speaker 0>pretty indicative following our last, our last article

00:09:42.900 --> 00:09:45.140
<v Speaker 0>or our lead story. So a little bit

00:09:45.140 --> 00:09:48.500
<v Speaker 0>of context. An anonymous employee claimed that Amazon

00:09:48.500 --> 00:09:50.340
<v Speaker 0>had laid off nearly half of its DevOps

00:09:50.340 --> 00:09:52.745
<v Speaker 0>teams at AWS and replaced them with AI

00:09:52.745 --> 00:09:56.045
<v Speaker 0>automation just days before a significant outage disrupted

00:09:56.105 --> 00:09:59.785
<v Speaker 0>major clients. Amazon denies the timing was related,

00:09:59.785 --> 00:10:02.905
<v Speaker 0>but internal Slack messages suggest otherwise. Oh, yeah.

00:10:02.905 --> 00:10:07.250
<v Speaker 0>It's always the Slack messages. And then, you

00:10:07.250 --> 00:10:09.410
<v Speaker 0>know, what does this kinda mean for us?

00:10:09.410 --> 00:10:10.850
<v Speaker 0>If this is true, I feel like this

00:10:10.850 --> 00:10:13.089
<v Speaker 0>is, like, a glimpse at the near term

00:10:13.089 --> 00:10:17.970
<v Speaker 0>risks of replacing, like, critical human infrastructure roles

00:10:17.970 --> 00:10:20.450
<v Speaker 0>with automation that's, like, not proven. You know?

00:10:20.450 --> 00:10:24.585
<v Speaker 0>It's it's it's not tested with time. These

00:10:24.585 --> 00:10:26.905
<v Speaker 0>employees are are dynamic and are able to

00:10:26.905 --> 00:10:29.625
<v Speaker 0>make decisions to fix things on the fly,

00:10:29.625 --> 00:10:35.600
<v Speaker 0>whereas automation tools aren't as flexible. And the

00:10:35.600 --> 00:10:38.339
<v Speaker 0>question really isn't whether AI can do DevOps.

00:10:38.399 --> 00:10:40.800
<v Speaker 0>It's whether, like, it should, especially without, like,

00:10:40.800 --> 00:10:42.880
<v Speaker 0>any sort of redundancy or fail safes in

00:10:42.880 --> 00:10:46.160
<v Speaker 0>place. But outside of that, let's just get

00:10:46.160 --> 00:10:49.555
<v Speaker 0>into the article and see, like, what they've

00:10:49.555 --> 00:10:54.834
<v Speaker 0>got for us. And, yeah, I mean, it's

00:10:54.995 --> 00:10:57.495
<v Speaker 0>so there's articles outlining some of the platforms

00:10:57.635 --> 00:11:00.435
<v Speaker 0>that were affected by this AWS outage, you

00:11:00.435 --> 00:11:06.380
<v Speaker 0>know, Snapchat, McDonald, Roblox, Fortnite, you name it.

00:11:06.380 --> 00:11:08.320
<v Speaker 0>A lot of people were down that day.

00:11:08.940 --> 00:11:10.540
<v Speaker 0>Servers are back up at this point, but

00:11:10.540 --> 00:11:12.300
<v Speaker 0>it was discovered that an interesting report had

00:11:12.300 --> 00:11:14.620
<v Speaker 0>been published right before the outage, alleging that

00:11:14.620 --> 00:11:16.300
<v Speaker 0>the company had laid off forty percent of

00:11:16.300 --> 00:11:18.640
<v Speaker 0>its DevOps teams to replace them with AI.

00:11:18.995 --> 00:11:20.755
<v Speaker 0>The author says that there was an email

00:11:20.755 --> 00:11:23.154
<v Speaker 0>memo, which was briefly posted on the internal

00:11:23.154 --> 00:11:25.795
<v Speaker 0>Wiki before being taken down, blamed the cuts

00:11:25.795 --> 00:11:30.774
<v Speaker 0>on strategic automation initiatives. It reports that AI

00:11:31.040 --> 00:11:34.560
<v Speaker 0>detects and fixes I'm permissions errors instantly, reveals

00:11:34.560 --> 00:11:37.280
<v Speaker 0>broken VPC or subnet conflicts, and rolls back

00:11:37.280 --> 00:11:40.480
<v Speaker 0>failed Lambda deployments without human input. Now there's

00:11:40.480 --> 00:11:42.160
<v Speaker 0>a lot of skepticism around this article, and

00:11:42.160 --> 00:11:43.440
<v Speaker 0>you should take it with a huge grain

00:11:43.440 --> 00:11:45.120
<v Speaker 0>of salt. But the timing is curious, although

00:11:45.120 --> 00:11:46.435
<v Speaker 0>we do not claim it is true or

00:11:46.435 --> 00:11:48.355
<v Speaker 0>somehow connected to the crash. Okay. So this

00:11:48.355 --> 00:11:51.654
<v Speaker 0>is all sort of hearsay. Nothing really, like,

00:11:51.954 --> 00:11:54.595
<v Speaker 0>confirmed, confirmed. But regardless, I mean, to hear

00:11:54.595 --> 00:11:56.995
<v Speaker 0>some of that, like, internal buzz or, like,

00:11:56.995 --> 00:12:00.780
<v Speaker 0>get that you know, those inklings of what

00:12:00.780 --> 00:12:02.860
<v Speaker 0>could be going on here, it's kinda really

00:12:02.860 --> 00:12:06.320
<v Speaker 0>interesting, to piece together. So I feel like,

00:12:09.820 --> 00:12:11.655
<v Speaker 0>you know, who knows if this is true,

00:12:11.655 --> 00:12:13.995
<v Speaker 0>whether it's forty percent or twenty percent regardless?

00:12:14.055 --> 00:12:17.435
<v Speaker 0>I'm sure that some automation changes have negatively

00:12:17.655 --> 00:12:20.935
<v Speaker 0>impacted, you know, the service in such a

00:12:20.935 --> 00:12:22.775
<v Speaker 0>way that it led to this crash. I'm

00:12:22.775 --> 00:12:26.030
<v Speaker 0>sure. You know? But this article is from

00:12:26.030 --> 00:12:29.150
<v Speaker 0>eighty dot l v. Thank you guys so

00:12:29.150 --> 00:12:35.410
<v Speaker 0>much for sharing. But really interesting, article from

00:12:36.030 --> 00:12:38.610
<v Speaker 0>eighty dot l v today just kind of,

00:12:41.235 --> 00:12:44.595
<v Speaker 0>almost like a great follow-up to that previous

00:12:44.595 --> 00:12:47.075
<v Speaker 0>one, like, whether we should slow down AI

00:12:47.075 --> 00:12:49.815
<v Speaker 0>changes. Like, are we changing are we removing

00:12:49.955 --> 00:12:52.755
<v Speaker 0>too many people? I saw another article earlier

00:12:52.755 --> 00:12:56.230
<v Speaker 0>that was, like, saying, like, by implementing AI,

00:12:56.230 --> 00:12:59.210
<v Speaker 0>we can remove jobs that are, like, pointless,

00:12:59.350 --> 00:13:01.270
<v Speaker 0>but we wouldn't have those jobs if they

00:13:01.270 --> 00:13:03.350
<v Speaker 0>were pointless. You know? It it was it

00:13:03.350 --> 00:13:04.790
<v Speaker 0>came up for a reason. I'm sure that

00:13:04.790 --> 00:13:06.550
<v Speaker 0>there is some sort of waste and that

00:13:06.550 --> 00:13:08.985
<v Speaker 0>there are some jobs that should have some

00:13:08.985 --> 00:13:12.505
<v Speaker 0>more overlap with other positions, but it's not

00:13:12.505 --> 00:13:14.825
<v Speaker 0>necessarily true that it's waste all the time.

00:13:14.825 --> 00:13:16.825
<v Speaker 0>So I wonder I wonder where that's driven

00:13:16.825 --> 00:13:19.545
<v Speaker 0>from. Would you trust an AI to manage

00:13:19.545 --> 00:13:23.660
<v Speaker 0>uptime for your favorite services? Would, like, would

00:13:25.240 --> 00:13:27.000
<v Speaker 0>would you trust AI for DevOps? I feel

00:13:27.000 --> 00:13:29.080
<v Speaker 0>like an AI and, you know, a a

00:13:29.080 --> 00:13:32.280
<v Speaker 0>human that teamed up could be good, you

00:13:32.280 --> 00:13:34.840
<v Speaker 0>know, just like how we're doing with all

00:13:34.840 --> 00:13:39.185
<v Speaker 0>the code gen nowadays, but, you know, I

00:13:39.185 --> 00:13:40.865
<v Speaker 0>don't know. I don't feel like it's ready

00:13:40.865 --> 00:13:45.985
<v Speaker 0>for unsupervised free reign in those, you know,

00:13:45.985 --> 00:13:52.000
<v Speaker 0>critical infrastructures, which nowadays, services like AWS is

00:13:52.000 --> 00:13:56.480
<v Speaker 0>a critical infrastructure. It's crazy to say. But,

00:13:56.480 --> 00:13:59.780
<v Speaker 0>yeah, let's move on. I found this article

00:13:59.840 --> 00:14:02.805
<v Speaker 0>on Medium. I don't know I know they're

00:14:02.805 --> 00:14:06.805
<v Speaker 0>kinda sticky with sharing sometimes, so hopefully, there's

00:14:06.805 --> 00:14:10.805
<v Speaker 0>no issue with, checking this article out for

00:14:10.805 --> 00:14:12.325
<v Speaker 0>you guys later. But let me just give

00:14:12.325 --> 00:14:13.925
<v Speaker 0>you a little rundown to give you an

00:14:13.925 --> 00:14:17.760
<v Speaker 0>idea of what they're sharing. Some context here.

00:14:17.760 --> 00:14:22.500
<v Speaker 0>Microsoft's new gaming Copilot, faces a privacy backlash.

00:14:23.520 --> 00:14:25.700
<v Speaker 0>So if you guys don't know what Copilot

00:14:25.760 --> 00:14:28.880
<v Speaker 0>is, it's that it's Microsoft's, like, AI that's

00:14:28.880 --> 00:14:35.785
<v Speaker 0>built into their Windows application. It's I feel

00:14:35.785 --> 00:14:37.725
<v Speaker 0>like it's one of the most negative experiences

00:14:38.025 --> 00:14:40.445
<v Speaker 0>about logging into a Windows PC or experiencing

00:14:40.505 --> 00:14:43.305
<v Speaker 0>Windows update is that Copilot is is there,

00:14:43.305 --> 00:14:45.305
<v Speaker 0>and you have to try to remove or

00:14:45.305 --> 00:14:47.620
<v Speaker 0>hide it. I I feel like I had

00:14:47.760 --> 00:14:49.760
<v Speaker 0>issues removing it the first time when I

00:14:49.760 --> 00:14:52.800
<v Speaker 0>had updated upgraded my, my old PC to

00:14:52.800 --> 00:14:55.920
<v Speaker 0>Windows eleven, and that just sent me off.

00:14:55.920 --> 00:14:58.960
<v Speaker 0>I literally uninstalled Windows at that point. It

00:14:58.960 --> 00:15:01.120
<v Speaker 0>switched my thing over to full Linux. So

00:15:01.120 --> 00:15:03.265
<v Speaker 0>it's, I mean, there's some issues with that

00:15:03.265 --> 00:15:07.425
<v Speaker 0>too, but I don't know. It just the

00:15:07.425 --> 00:15:12.545
<v Speaker 0>layers of, you know, privacy invasion that's going

00:15:12.545 --> 00:15:15.605
<v Speaker 0>on is crazy. But here's a new one.

00:15:16.290 --> 00:15:19.010
<v Speaker 0>Microsoft new gaming copilot, an AI system built

00:15:19.010 --> 00:15:21.970
<v Speaker 0>into Windows recall and Xbox Game Bar, has

00:15:21.970 --> 00:15:25.350
<v Speaker 0>already failed multiple independent private privacy audits. Researchers

00:15:25.410 --> 00:15:28.529
<v Speaker 0>found it recording desktop screenshots and chat logs

00:15:28.529 --> 00:15:35.305
<v Speaker 0>even after users disabled data learning features. You

00:15:35.305 --> 00:15:38.185
<v Speaker 0>know, after the recall controversy, this reignites the

00:15:38.185 --> 00:15:41.785
<v Speaker 0>debate over Microsoft's growing surveillance footprint inside of

00:15:41.785 --> 00:15:44.985
<v Speaker 0>Windows. If privacy settings don't actually work, users

00:15:44.985 --> 00:15:47.465
<v Speaker 0>may rethink how tightly integrated AI should be

00:15:47.465 --> 00:15:50.740
<v Speaker 0>with their games and operating systems. Yeah. Is

00:15:50.740 --> 00:15:53.300
<v Speaker 0>convenience worth the cost of surveillance even in

00:15:53.300 --> 00:15:57.140
<v Speaker 0>gaming? You know? That's that's ridiculous. I I

00:15:57.140 --> 00:15:59.140
<v Speaker 0>was laughing at this picture that, you know,

00:15:59.140 --> 00:16:03.295
<v Speaker 0>it was like a canonically accurate cookies message,

00:16:03.295 --> 00:16:05.714
<v Speaker 0>and it was just like, you can reject

00:16:06.014 --> 00:16:07.855
<v Speaker 0>the cookies, but we're gonna collect your data

00:16:07.855 --> 00:16:09.535
<v Speaker 0>anyway. And I was just cracking up to

00:16:09.535 --> 00:16:14.750
<v Speaker 0>say, yeah. That's the truth. But here's the

00:16:14.750 --> 00:16:17.630
<v Speaker 0>article. With Windows eleven, Microsoft are currently rolling

00:16:17.630 --> 00:16:20.510
<v Speaker 0>out gaming Copilot to all customers. This uses

00:16:20.510 --> 00:16:23.149
<v Speaker 0>Copilot vision similar to Recall Under the Hood

00:16:23.149 --> 00:16:25.550
<v Speaker 0>and basically inspects what gamers are doing using

00:16:25.550 --> 00:16:27.390
<v Speaker 0>AI in order to give them hints and

00:16:27.390 --> 00:16:30.565
<v Speaker 0>tips and such. The software is labeled beta,

00:16:30.565 --> 00:16:32.725
<v Speaker 0>but is being pushed out to existing PCs

00:16:32.725 --> 00:16:35.765
<v Speaker 0>anyway because, well, why not beta something in

00:16:35.765 --> 00:16:38.725
<v Speaker 0>prod? It's twenty twenty five after all. I

00:16:38.725 --> 00:16:42.390
<v Speaker 0>hate this. I hate this. It it appears

00:16:42.390 --> 00:16:44.970
<v Speaker 0>in game bar, which is accessible from Windows

00:16:45.110 --> 00:16:47.510
<v Speaker 0>key plus g, and you may not have

00:16:47.510 --> 00:16:52.070
<v Speaker 0>it yet as they are staging deployment. So

00:16:52.070 --> 00:16:53.990
<v Speaker 0>if you even if you're not in Windows

00:16:53.990 --> 00:16:55.830
<v Speaker 0>install insiders, it looks like you'll have it

00:16:55.830 --> 00:17:03.265
<v Speaker 0>installed. Yeah. And even if you uninstall Copilot

00:17:03.265 --> 00:17:06.325
<v Speaker 0>from your PC, this gaming Copilot gets silently

00:17:06.385 --> 00:17:10.169
<v Speaker 0>installed anyway. So that's unfair and ridiculous. I

00:17:10.169 --> 00:17:12.030
<v Speaker 0>feel like that should be a trickle down

00:17:12.250 --> 00:17:17.049
<v Speaker 0>no. Yeah. No onboarding, no notification that it's

00:17:17.049 --> 00:17:20.890
<v Speaker 0>installed, and no privacy options. And the writer's

00:17:20.890 --> 00:17:23.150
<v Speaker 0>in the UK, so they're sharing that this

00:17:23.825 --> 00:17:25.825
<v Speaker 0>kinda shocked them because they've they've been pretty

00:17:25.825 --> 00:17:30.225
<v Speaker 0>critical on privacy as you may know. But

00:17:30.225 --> 00:17:33.525
<v Speaker 0>definitely check this article out if you're I

00:17:33.825 --> 00:17:35.265
<v Speaker 0>wonder if there is actually a way to

00:17:35.265 --> 00:17:41.240
<v Speaker 0>turn it off. Let me see. Okay. You

00:17:41.240 --> 00:17:43.080
<v Speaker 0>can disable the gaming Copilot widget from the

00:17:43.080 --> 00:17:46.040
<v Speaker 0>game bar. However, from experimentation so far, the

00:17:46.040 --> 00:17:48.200
<v Speaker 0>network traffic is still running with the user

00:17:48.200 --> 00:17:50.280
<v Speaker 0>interface not open. I don't know if this

00:17:50.280 --> 00:17:59.625
<v Speaker 0>is a feature or a bug. Well, regardless,

00:17:59.684 --> 00:18:01.125
<v Speaker 0>I would recommend turning it off if you're

00:18:01.205 --> 00:18:03.044
<v Speaker 0>if you don't plan on using it or

00:18:03.044 --> 00:18:04.245
<v Speaker 0>if you don't want your data to be

00:18:04.245 --> 00:18:06.840
<v Speaker 0>sent to Microsoft. You can use it to

00:18:06.840 --> 00:18:08.440
<v Speaker 0>help you out with your your game time,

00:18:08.440 --> 00:18:12.380
<v Speaker 0>but then then what's the point then of

00:18:12.600 --> 00:18:14.120
<v Speaker 0>of trying to, like, you know, crack the

00:18:14.120 --> 00:18:17.640
<v Speaker 0>code in the game or whatever? Maybe if

00:18:17.640 --> 00:18:19.320
<v Speaker 0>it's, like, map hints or something for people

00:18:19.320 --> 00:18:21.655
<v Speaker 0>that are unfamiliar with game mechanics, I could

00:18:21.655 --> 00:18:29.415
<v Speaker 0>see that. Alright. Well, why don't we move

00:18:29.415 --> 00:18:33.335
<v Speaker 0>on to the next article? This one is

00:18:33.335 --> 00:18:40.500
<v Speaker 0>from VentureBeat. Is this from where is this?

00:18:41.120 --> 00:18:43.920
<v Speaker 0>I think I did wanna add this article

00:18:43.920 --> 00:18:46.020
<v Speaker 0>because I thought this was kinda sweet for

00:18:46.240 --> 00:18:51.565
<v Speaker 0>all you, LLM users On the opposite end

00:18:51.565 --> 00:18:53.644
<v Speaker 0>of the spectrum, I guess, from our previous

00:18:53.644 --> 00:18:57.585
<v Speaker 0>conversation, researchers discovered one simple prompt that boosts

00:18:57.644 --> 00:19:01.325
<v Speaker 0>AI accuracy. And AI researchers have found that

00:19:01.325 --> 00:19:03.884
<v Speaker 0>adding a single sentence, essentially a meta instruction

00:19:03.884 --> 00:19:06.350
<v Speaker 0>for chain of thought reasoning, improves accuracy and

00:19:06.350 --> 00:19:09.230
<v Speaker 0>consistency across major language models by up to

00:19:09.230 --> 00:19:12.510
<v Speaker 0>thirty percent. The phrase triggers deeper reasoning steps

00:19:12.510 --> 00:19:17.390
<v Speaker 0>before producing an answer. And I think kind

00:19:17.390 --> 00:19:19.950
<v Speaker 0>of understanding how this phrase or this tool

00:19:19.950 --> 00:19:23.285
<v Speaker 0>works will help you understand just how, AI

00:19:23.285 --> 00:19:26.345
<v Speaker 0>tools actually work, that they're not actually responding

00:19:26.405 --> 00:19:27.685
<v Speaker 0>to you. You know what I mean? I

00:19:27.685 --> 00:19:30.805
<v Speaker 0>mean, in a sense. Why don't I read

00:19:30.805 --> 00:19:33.125
<v Speaker 0>the article and you guys can get a

00:19:33.125 --> 00:19:34.805
<v Speaker 0>sense for what I'm kinda trying to tell

00:19:34.805 --> 00:19:37.730
<v Speaker 0>you? One of the coolest things about generative

00:19:37.730 --> 00:19:40.450
<v Speaker 0>AI models, both large language models and diffusion

00:19:40.450 --> 00:19:42.950
<v Speaker 0>based image generators, is that they are nondeterministic.

00:19:44.210 --> 00:19:47.010
<v Speaker 0>This is the key word here, nondeterministic. It

00:19:47.010 --> 00:19:49.650
<v Speaker 0>is despite their reputation among some critics as

00:19:49.650 --> 00:19:53.784
<v Speaker 0>being fancy autocorrect, generative AI models actually generate

00:19:53.784 --> 00:19:56.184
<v Speaker 0>their outlets by choosing from a distribution of

00:19:56.184 --> 00:20:00.125
<v Speaker 0>the most probable NEX tokens, units of information,

00:20:00.664 --> 00:20:05.929
<v Speaker 0>to fill out their response. Asking an LLM,

00:20:05.990 --> 00:20:08.070
<v Speaker 0>what is the capital of France, will have

00:20:08.070 --> 00:20:11.289
<v Speaker 0>have it sample its probability of probability distribution

00:20:11.590 --> 00:20:14.549
<v Speaker 0>for France, capital cities, etcetera, to arrive at

00:20:14.549 --> 00:20:17.190
<v Speaker 0>the answer, Paris. But the answer could come

00:20:17.190 --> 00:20:19.055
<v Speaker 0>in the format, the capital of France is

00:20:19.055 --> 00:20:21.775
<v Speaker 0>Paris or simply Paris or Paris, though it

00:20:21.775 --> 00:20:24.495
<v Speaker 0>was Versailles at one point. So you can

00:20:24.495 --> 00:20:28.175
<v Speaker 0>kinda see it's based on pattern matching rather

00:20:28.175 --> 00:20:33.990
<v Speaker 0>than, like, thinking of the response, which maybe

00:20:33.990 --> 00:20:36.310
<v Speaker 0>some people argue there's not much difference there,

00:20:36.310 --> 00:20:46.245
<v Speaker 0>but but that understanding this this this prompt

00:20:46.245 --> 00:20:48.905
<v Speaker 0>will will will kind of show you, the

00:20:48.965 --> 00:20:50.645
<v Speaker 0>the term that they tell you to use

00:20:50.645 --> 00:20:54.585
<v Speaker 0>here. So it's gonna be generate five responses

00:20:54.725 --> 00:20:58.085
<v Speaker 0>with their corresponding probabilities sampled from the full

00:20:58.085 --> 00:21:02.480
<v Speaker 0>distribution. And this method, it's called verbalized sampling,

00:21:02.779 --> 00:21:04.779
<v Speaker 0>helps models like g p t four, Claude,

00:21:04.779 --> 00:21:07.419
<v Speaker 0>and Gemini produce more diverse and human like

00:21:07.419 --> 00:21:11.200
<v Speaker 0>outputs without retraining or access to internal parameters,

00:21:14.345 --> 00:21:18.025
<v Speaker 0>which is kinda cool. LLMs so this is

00:21:18.025 --> 00:21:19.545
<v Speaker 0>a quote from the the person who wrote

00:21:19.545 --> 00:21:22.105
<v Speaker 0>the paper, Wei Yan Shi, an assistant professor

00:21:22.105 --> 00:21:25.865
<v Speaker 0>at Northeastern University. So they're one of the

00:21:25.865 --> 00:21:29.950
<v Speaker 0>coauthors. LLMs' potentials are not fully unlocked yet,

00:21:29.950 --> 00:21:32.270
<v Speaker 0>as shown in our paper. Prompt optimization can

00:21:32.270 --> 00:21:34.510
<v Speaker 0>be guided by thinking how about how LLMs

00:21:34.510 --> 00:21:36.910
<v Speaker 0>are trained and aligned and can be proved

00:21:36.910 --> 00:21:42.235
<v Speaker 0>theoretically. So then they kinda get into model

00:21:42.235 --> 00:21:46.875
<v Speaker 0>collapse and how this verbalized sampling reverses model

00:21:47.035 --> 00:21:49.595
<v Speaker 0>or model collapse. So the root cause of

00:21:49.595 --> 00:21:51.915
<v Speaker 0>model collapse lies not just in algorithms like

00:21:51.915 --> 00:21:54.330
<v Speaker 0>reinforcement learning from human feedback, but in the

00:21:54.330 --> 00:21:57.130
<v Speaker 0>structure of human preferences. People tend to rate

00:21:57.130 --> 00:22:00.110
<v Speaker 0>more familiar or typical answers as better, which

00:22:00.170 --> 00:22:03.610
<v Speaker 0>nudges LLMs towards safe choices over diverse ones

00:22:03.610 --> 00:22:07.290
<v Speaker 0>during fine tuning. However, this bias doesn't erase

00:22:07.290 --> 00:22:10.794
<v Speaker 0>models' underlying knowledge. It just suppresses it. Versus

00:22:10.794 --> 00:22:14.315
<v Speaker 0>works by bypassing this suppression. Instead of asking

00:22:14.315 --> 00:22:16.975
<v Speaker 0>for the single most likely output, it invites

00:22:17.034 --> 00:22:19.534
<v Speaker 0>the model to reveal a set of plausible

00:22:19.595 --> 00:22:23.679
<v Speaker 0>responses and their relative probabilities. This distribution level

00:22:23.679 --> 00:22:26.799
<v Speaker 0>prompting restores access to the richer diversity present

00:22:26.799 --> 00:22:32.480
<v Speaker 0>in the base per pretraining model. So that's

00:22:32.480 --> 00:22:35.755
<v Speaker 0>kinda cool. Try this out. I I will

00:22:35.755 --> 00:22:37.675
<v Speaker 0>include the article in the show notes as

00:22:37.675 --> 00:22:39.434
<v Speaker 0>well as the newsletter so you guys can

00:22:39.434 --> 00:22:42.335
<v Speaker 0>get a little bit more detail from, VentureBeat.

00:22:42.475 --> 00:22:46.395
<v Speaker 0>This was written by Carl Franzen, but as

00:22:46.395 --> 00:22:49.535
<v Speaker 0>you guys can tell, like, it's based on

00:22:50.460 --> 00:22:56.299
<v Speaker 0>probability oh, sorry. Probability, for to provide you,

00:22:56.540 --> 00:22:58.380
<v Speaker 0>a context, you know, based on, you know,

00:22:58.380 --> 00:23:01.100
<v Speaker 0>what's the most likely response to the respond

00:23:01.500 --> 00:23:05.085
<v Speaker 0>the the prompt that you provided it. So

00:23:05.085 --> 00:23:07.165
<v Speaker 0>asking it for multiple, you know, along, like,

00:23:07.165 --> 00:23:09.725
<v Speaker 0>the most likely vectors or or whatever the

00:23:09.725 --> 00:23:11.165
<v Speaker 0>statement was. One more time. Let's do it.

00:23:11.165 --> 00:23:12.605
<v Speaker 0>Let's say the statement one more time so

00:23:12.605 --> 00:23:15.565
<v Speaker 0>we don't lose it. Generate five responses with

00:23:15.565 --> 00:23:19.345
<v Speaker 0>their corresponding probabilities sampled from the full distribution.

00:23:20.120 --> 00:23:24.040
<v Speaker 0>Alright. So, yeah. And this will help with

00:23:24.040 --> 00:23:26.380
<v Speaker 0>model collapse, help you get more diverse answers

00:23:26.440 --> 00:23:29.740
<v Speaker 0>in your prompts, if you are using LLMs,

00:23:31.400 --> 00:23:33.255
<v Speaker 0>which is I I think that's pretty cool.

00:23:33.655 --> 00:23:36.215
<v Speaker 0>Prompt optimization really isn't talked about all that

00:23:36.215 --> 00:23:39.655
<v Speaker 0>much, I feel like, and I think this

00:23:39.655 --> 00:23:42.315
<v Speaker 0>is where you're getting the people that are

00:23:42.535 --> 00:23:45.815
<v Speaker 0>optimizing their prompts are getting the best out

00:23:45.815 --> 00:23:50.690
<v Speaker 0>of their LLM usage. You know? They're not

00:23:50.690 --> 00:23:52.529
<v Speaker 0>being led by the nose, by the the

00:23:52.529 --> 00:23:55.029
<v Speaker 0>model, and instead are able to make decisions

00:23:55.169 --> 00:23:59.169
<v Speaker 0>based on, available options or routes that you

00:23:59.169 --> 00:24:02.075
<v Speaker 0>could take. So I could see now, you

00:24:02.075 --> 00:24:04.715
<v Speaker 0>know, looking back how this might, change how

00:24:04.715 --> 00:24:06.815
<v Speaker 0>I might have prompted something in the past.

00:24:07.915 --> 00:24:10.875
<v Speaker 0>So kinda cool to think about, kinda cool

00:24:10.875 --> 00:24:14.279
<v Speaker 0>to look at. But definitely look into this

00:24:14.279 --> 00:24:15.880
<v Speaker 0>or read this article if you guys get

00:24:15.880 --> 00:24:20.120
<v Speaker 0>the chance. There's a lot more options. I've

00:24:20.279 --> 00:24:23.240
<v Speaker 0>it looks like there's even a Python package

00:24:23.240 --> 00:24:25.720
<v Speaker 0>that allows you to do a verbalized sampling.

00:24:25.720 --> 00:24:28.625
<v Speaker 0>So if any of you are Python users

00:24:28.625 --> 00:24:30.385
<v Speaker 0>and would like to try that, pip install

00:24:30.385 --> 00:24:36.225
<v Speaker 0>verbalized sampling, check that out. But that's kinda

00:24:36.225 --> 00:24:40.005
<v Speaker 0>cool. Let's move on to the next article.

00:24:40.519 --> 00:24:43.559
<v Speaker 0>We already talked about Wozniak pushing back against

00:24:43.559 --> 00:24:47.799
<v Speaker 0>the AI superintelligence serve vibe coding fade fades

00:24:47.799 --> 00:24:51.080
<v Speaker 0>as Wozniak pushes for a practical successor. This

00:24:51.080 --> 00:24:53.320
<v Speaker 0>was from the new stack written by Daryl

00:24:53.320 --> 00:24:56.825
<v Speaker 0>Taft. Steve Wozniak is and this is just

00:24:56.825 --> 00:24:58.425
<v Speaker 0>some context before we get into the article.

00:24:58.425 --> 00:25:00.745
<v Speaker 0>Steve Wozniak has resurfaced in the dev world

00:25:00.745 --> 00:25:03.225
<v Speaker 0>promoting a new production ready coding framework as

00:25:03.225 --> 00:25:05.625
<v Speaker 0>a replacement for the short lived vibe coding

00:25:05.625 --> 00:25:08.445
<v Speaker 0>trend, an approach that emphasize flow and creativity

00:25:08.665 --> 00:25:11.809
<v Speaker 0>over structure. His alternative aims to merge intuition

00:25:11.870 --> 00:25:14.830
<v Speaker 0>with industrial grade reliability. Yeah. So it's kind

00:25:14.830 --> 00:25:16.030
<v Speaker 0>of one thing about vibe coding that I

00:25:16.030 --> 00:25:17.790
<v Speaker 0>don't understand is, like, what do you mean

00:25:17.790 --> 00:25:21.470
<v Speaker 0>creativity over structure? Are we not I understand

00:25:21.470 --> 00:25:25.125
<v Speaker 0>that as the vibe coder, quote, unquote, you

00:25:25.125 --> 00:25:27.924
<v Speaker 0>don't interact with the quote, code question mark.

00:25:27.924 --> 00:25:31.385
<v Speaker 0>That's my my understanding of the whole process.

00:25:31.525 --> 00:25:33.285
<v Speaker 0>So you don't really care about the structure

00:25:33.285 --> 00:25:37.140
<v Speaker 0>of the code, repeatability. I mean, I feel

00:25:37.140 --> 00:25:39.060
<v Speaker 0>like my issue with it is the the

00:25:39.060 --> 00:25:41.560
<v Speaker 0>the MLM isn't thinking of, oh, best security

00:25:41.620 --> 00:25:44.200
<v Speaker 0>practices for you. The MLM isn't thinking of

00:25:44.740 --> 00:25:49.424
<v Speaker 0>reusing code or or optimizing, any of your

00:25:49.424 --> 00:25:51.264
<v Speaker 0>pipelines. You know what I mean? It's not

00:25:51.264 --> 00:25:53.024
<v Speaker 0>thinking about that unless you ask it to

00:25:53.024 --> 00:25:55.345
<v Speaker 0>think that or ask it to to do

00:25:55.345 --> 00:25:59.024
<v Speaker 0>that. So kinda I just wonder how Vibe

00:25:59.024 --> 00:26:01.345
<v Speaker 0>coders have been and I use this word

00:26:01.345 --> 00:26:05.630
<v Speaker 0>lightly successful, quote, unquote successful. There must be

00:26:05.630 --> 00:26:08.690
<v Speaker 0>some sort of technique they're using. But Wozniak's

00:26:08.830 --> 00:26:12.130
<v Speaker 0>comeback reflects a growing frustration among developers. Creativity

00:26:12.269 --> 00:26:15.789
<v Speaker 0>matters, but shipping stable software still wins. His

00:26:15.789 --> 00:26:18.669
<v Speaker 0>message? Tools should serve humans, not the other

00:26:18.669 --> 00:26:23.174
<v Speaker 0>way around. Yes. So Boz launches to build

00:26:23.235 --> 00:26:25.475
<v Speaker 0>production ready apps using pre built components and

00:26:25.475 --> 00:26:27.794
<v Speaker 0>human oversight as traffic leads to vibe coding

00:26:27.794 --> 00:26:31.414
<v Speaker 0>tools plummets up to sixty four percent. Traffic

00:26:31.554 --> 00:26:34.355
<v Speaker 0>to leading vibe coding tools plummets up to

00:26:34.355 --> 00:26:38.130
<v Speaker 0>sixty four percent. Wow. What what is what

00:26:38.130 --> 00:26:42.470
<v Speaker 0>led to this? Did I miss something? Okay.

00:26:42.610 --> 00:26:44.529
<v Speaker 0>Ben Collins and Brad Eckert say that they

00:26:44.529 --> 00:26:46.770
<v Speaker 0>have been watching the app building bubble deflate

00:26:46.770 --> 00:26:49.169
<v Speaker 0>in real time. It's been re it's been

00:26:49.169 --> 00:26:51.395
<v Speaker 0>pretty dramatic, said Collins, who is cofounder and

00:26:51.395 --> 00:26:54.995
<v Speaker 0>CEO of new AI platform startup, Woz. Since

00:26:54.995 --> 00:26:58.375
<v Speaker 0>June, traffic to Lovable has dropped forty percent.

00:26:58.995 --> 00:27:01.635
<v Speaker 0>Bolt dot news down twenty percent twenty seven

00:27:01.635 --> 00:27:04.915
<v Speaker 0>percent. Vercel's v zero down sixty four percent

00:27:04.915 --> 00:27:07.429
<v Speaker 0>since May, all according to data from Barclays.

00:27:08.050 --> 00:27:10.550
<v Speaker 0>The two I the two MIT trained engineers

00:27:10.610 --> 00:27:12.290
<v Speaker 0>who were roommates over a decade ago think

00:27:12.290 --> 00:27:14.610
<v Speaker 0>they know why. Those tools were never meant

00:27:14.610 --> 00:27:16.770
<v Speaker 0>to build real products. They were built to

00:27:16.770 --> 00:27:21.215
<v Speaker 0>create demos. Everyone got really excited about going

00:27:21.215 --> 00:27:23.394
<v Speaker 0>from prompt to a working prototype in minutes.

00:27:23.774 --> 00:27:25.774
<v Speaker 0>But then the reality sets in. You got

00:27:25.774 --> 00:27:28.174
<v Speaker 0>this mess of code, security holes, and no

00:27:28.174 --> 00:27:31.890
<v Speaker 0>way to actually maintain it. Wow. Yeah. It's

00:27:32.050 --> 00:27:33.810
<v Speaker 0>almost like you don't interact with the code

00:27:33.810 --> 00:27:39.010
<v Speaker 0>when you're vibe coding or something. So I

00:27:39.010 --> 00:27:41.090
<v Speaker 0>guess am I being baited? Is this a

00:27:41.090 --> 00:27:45.410
<v Speaker 0>trick? I guess it has nothing to do

00:27:45.410 --> 00:27:49.755
<v Speaker 0>with Wozniak. No, it does, Steve Wozniak. Okay.

00:27:49.755 --> 00:27:51.515
<v Speaker 0>Let's just keep going. Maybe we'll figure this

00:27:51.515 --> 00:27:57.355
<v Speaker 0>out. So they built San Francisco based Woz,

00:27:57.355 --> 00:27:59.195
<v Speaker 0>which launched last week with six million in

00:27:59.195 --> 00:28:01.460
<v Speaker 0>funding to go on automation with actual human

00:28:01.460 --> 00:28:03.800
<v Speaker 0>engineers to build apps that go beyond prototypes.

00:28:05.540 --> 00:28:07.060
<v Speaker 0>Ask any engineer who has worked in a

00:28:07.060 --> 00:28:08.420
<v Speaker 0>big tech company and they'll tell you a

00:28:08.420 --> 00:28:10.980
<v Speaker 0>similar story. A product manager uses one of

00:28:10.980 --> 00:28:13.540
<v Speaker 0>these rapid prototyping tools, build something cool and

00:28:13.540 --> 00:28:16.274
<v Speaker 0>shows it to engineering. Engineering says, great. We

00:28:16.274 --> 00:28:18.195
<v Speaker 0>understand what you want, and then throws the

00:28:18.195 --> 00:28:20.674
<v Speaker 0>whole thing in the trash. Yes. It's exactly

00:28:20.674 --> 00:28:23.174
<v Speaker 0>what you do. They start it from scratch

00:28:23.554 --> 00:28:25.395
<v Speaker 0>because the code is unusable. It's what we

00:28:25.395 --> 00:28:30.395
<v Speaker 0>call spaghetti code. Everything is tangled together. And

00:28:30.395 --> 00:28:30.441
<v Speaker 0>you can't tell what's doing what and it's

00:28:30.441 --> 00:28:30.870
<v Speaker 0>full of shortcuts that will come back to

00:28:30.870 --> 00:28:34.710
<v Speaker 0>bite you later. Yes. Clean your code up.

00:28:34.710 --> 00:28:40.710
<v Speaker 0>Use templates, please. Or functions, you know, the

00:28:40.710 --> 00:28:44.775
<v Speaker 0>whole nine yards. That works fine if you're

00:28:44.775 --> 00:28:46.695
<v Speaker 0>Google or Meta and you have hundreds of

00:28:46.695 --> 00:28:49.655
<v Speaker 0>engineers, but most businesses don't. A restaurant chain

00:28:49.655 --> 00:28:52.215
<v Speaker 0>doesn't have an engineering team. A consulting firm

00:28:52.215 --> 00:28:53.975
<v Speaker 0>doesn't have an engineering team. They need the

00:28:53.975 --> 00:28:55.815
<v Speaker 0>final product, not a product they have to

00:28:55.815 --> 00:29:03.840
<v Speaker 0>spend six months rebuilding. So what they did

00:29:03.840 --> 00:29:06.740
<v Speaker 0>was they built a assembly line for apps.

00:29:09.039 --> 00:29:12.000
<v Speaker 0>WaaS system uses, prebuilt components as building blocks,

00:29:12.000 --> 00:29:14.175
<v Speaker 0>combining them based on what customer needs. When

00:29:14.175 --> 00:29:16.655
<v Speaker 0>something complicated comes up, an actual human engineer

00:29:16.655 --> 00:29:18.575
<v Speaker 0>steps in. And before anything goes live, a

00:29:18.575 --> 00:29:26.675
<v Speaker 0>human reviews the whole thing. So does this

00:29:27.559 --> 00:29:29.799
<v Speaker 0>so I guess they're using almost like a

00:29:29.799 --> 00:29:32.600
<v Speaker 0>factory method where they take components of, like,

00:29:32.600 --> 00:29:36.120
<v Speaker 0>websites and then automate that build, and then

00:29:36.120 --> 00:29:38.120
<v Speaker 0>someone goes and reviews it after. And most

00:29:38.120 --> 00:29:42.245
<v Speaker 0>of these components should be, you know, airtight

00:29:42.245 --> 00:29:48.725
<v Speaker 0>in terms of security practices and, I guess

00:29:48.725 --> 00:29:50.485
<v Speaker 0>that would be the main thing, security practices

00:29:50.485 --> 00:29:52.245
<v Speaker 0>and being able to be, like, readable and

00:29:52.245 --> 00:29:54.820
<v Speaker 0>whatnot. So what you're maintaining are more of

00:29:54.820 --> 00:29:58.120
<v Speaker 0>the templates rather than the actual factory. But

00:30:03.059 --> 00:30:04.659
<v Speaker 0>definitely check this article out if you want

00:30:04.659 --> 00:30:06.740
<v Speaker 0>a little more insight on this tool. Let

00:30:06.740 --> 00:30:08.760
<v Speaker 0>me see if I can find something. Woz

00:30:09.055 --> 00:30:17.295
<v Speaker 0>Woz platform. Woz, Woz I didn't realize how

00:30:17.295 --> 00:30:19.955
<v Speaker 0>little context that I was getting from that.

00:30:23.320 --> 00:30:25.560
<v Speaker 0>Okay. So it is a it's still an

00:30:25.560 --> 00:30:29.160
<v Speaker 0>AI platform. Yes. But in it but instead

00:30:29.160 --> 00:30:31.320
<v Speaker 0>of using writing the the code from scratch,

00:30:31.320 --> 00:30:33.640
<v Speaker 0>it's using components that are prebuilt. That's all

00:30:33.640 --> 00:30:35.775
<v Speaker 0>it is. So and this is just a

00:30:35.775 --> 00:30:37.635
<v Speaker 0>little bit cleaner. It's a little bit safer.

00:30:38.414 --> 00:30:40.495
<v Speaker 0>Just kinda reminds me of Legos, giving someone

00:30:40.495 --> 00:30:42.495
<v Speaker 0>Legos. Instead of letting them actually design the

00:30:42.495 --> 00:30:44.575
<v Speaker 0>Lego pieces, give them Lego pieces that work,

00:30:44.575 --> 00:30:47.210
<v Speaker 0>that can combine together nice and easy. So

00:30:47.210 --> 00:30:48.410
<v Speaker 0>I think that's a kind of a good

00:30:48.410 --> 00:30:50.970
<v Speaker 0>explanation for that. But definitely look into this

00:30:50.970 --> 00:30:53.850
<v Speaker 0>a little bit more. I'm sure something like

00:30:53.850 --> 00:30:57.930
<v Speaker 0>this will replace Vibe Coding, or or is

00:30:58.090 --> 00:30:59.985
<v Speaker 0>I don't think Vibe Coding is gonna work,

00:31:00.065 --> 00:31:03.025
<v Speaker 0>period. Definitely good for demoing, like they suggested

00:31:03.025 --> 00:31:05.825
<v Speaker 0>in the article. Like, a product project manager

00:31:05.825 --> 00:31:07.265
<v Speaker 0>comes up with a cool idea. It doesn't

00:31:07.265 --> 00:31:08.625
<v Speaker 0>know how to code. Let's make a quick

00:31:08.625 --> 00:31:12.145
<v Speaker 0>Vibe code demo. But, I just don't see

00:31:12.145 --> 00:31:18.010
<v Speaker 0>it being production grade anytime soon. Let's get

00:31:18.010 --> 00:31:20.330
<v Speaker 0>into the next article. This this this one

00:31:20.330 --> 00:31:22.570
<v Speaker 0>just caught me by the name, so I

00:31:22.570 --> 00:31:24.730
<v Speaker 0>hope you guys enjoyed it as well. Man

00:31:24.730 --> 00:31:28.315
<v Speaker 0>builds a two billion FPS camera that captures

00:31:28.375 --> 00:31:31.575
<v Speaker 0>light itself. This one's from Petapixel written by

00:31:31.575 --> 00:31:37.335
<v Speaker 0>Matt Grokut. Some context, engineer Andreas Velten has

00:31:37.335 --> 00:31:40.235
<v Speaker 0>created a camera capable of recording at two

00:31:40.295 --> 00:31:43.260
<v Speaker 0>billion frames per second fast enough to visualize

00:31:43.260 --> 00:31:46.700
<v Speaker 0>photons moving across objects. The project merges ultra

00:31:46.700 --> 00:31:49.900
<v Speaker 0>high speed imaging with computational reconstruction to literally

00:31:49.900 --> 00:31:53.660
<v Speaker 0>watch light bounce in real time. Beyond the

00:31:53.660 --> 00:31:56.140
<v Speaker 0>mind blowing visuals, this tech has major implications

00:31:56.140 --> 00:31:58.000
<v Speaker 0>for medical imaging, optical communication, and even AR

00:32:00.115 --> 00:32:03.095
<v Speaker 0>AR development. Seeing light in motion could redefine

00:32:03.155 --> 00:32:06.275
<v Speaker 0>how we model the physical world. What's the

00:32:06.275 --> 00:32:09.395
<v Speaker 0>next frontier once we can record reality at

00:32:09.395 --> 00:32:14.860
<v Speaker 0>light speed? What's yeah. What is beyond that?

00:32:15.080 --> 00:32:16.680
<v Speaker 0>I am super stoked for a lot of

00:32:16.680 --> 00:32:18.600
<v Speaker 0>the AR developments that should be coming up

00:32:18.600 --> 00:32:20.600
<v Speaker 0>in the next, you know, ten, fifteen years.

00:32:20.600 --> 00:32:23.080
<v Speaker 0>I think AR is cool. I think being

00:32:23.080 --> 00:32:27.895
<v Speaker 0>able to interact with, virtual reality in your

00:32:27.895 --> 00:32:30.295
<v Speaker 0>actual reality. So it's yeah. I guess, yeah,

00:32:30.295 --> 00:32:33.735
<v Speaker 0>augmented reality is is cool. I think that's

00:32:33.735 --> 00:32:36.875
<v Speaker 0>awesome. I wanna see more without massive headsets.

00:32:39.335 --> 00:32:42.789
<v Speaker 0>Let's get into the article. Brian Heidett of

00:32:42.789 --> 00:32:45.269
<v Speaker 0>the Alpha Phoenix YouTube channel has recorded a

00:32:45.269 --> 00:32:48.169
<v Speaker 0>laser beam at two billion frames per second,

00:32:48.630 --> 00:32:50.570
<v Speaker 0>recording the path of light as it bounces

00:32:50.630 --> 00:32:53.990
<v Speaker 0>across his garage. Heidett, who was a PhD

00:32:53.990 --> 00:32:56.355
<v Speaker 0>in material science, previously built a one billion

00:32:56.355 --> 00:32:59.155
<v Speaker 0>FPS camera and iterated on the previous design

00:32:59.155 --> 00:33:01.475
<v Speaker 0>to improve virtually everything over the course of

00:33:01.475 --> 00:33:04.275
<v Speaker 0>the last twelve months. The camera itself consists

00:33:04.275 --> 00:33:06.435
<v Speaker 0>of one mirror, one lens, two tubes, cables,

00:33:06.435 --> 00:33:08.355
<v Speaker 0>one of the weirdest camera flash bulbs probably

00:33:08.355 --> 00:33:10.355
<v Speaker 0>ever built, and a few hundred lines of

00:33:10.355 --> 00:33:13.610
<v Speaker 0>Python. It's only fast enough to watch light

00:33:13.610 --> 00:33:15.210
<v Speaker 0>move. It's fast enough to see the past,

00:33:15.210 --> 00:33:18.970
<v Speaker 0>exclaims the ebullient Hi Dad. Light moves about

00:33:18.970 --> 00:33:21.290
<v Speaker 0>six inches or fifteen centimeters per frame of

00:33:21.290 --> 00:33:23.450
<v Speaker 0>this video. The beam of light is traveling

00:33:23.450 --> 00:33:25.845
<v Speaker 0>at the universe's speed limit. Light in any

00:33:25.845 --> 00:33:27.925
<v Speaker 0>reference frame will never move any faster or

00:33:27.925 --> 00:33:36.805
<v Speaker 0>any slower than this speed. So I guess

00:33:36.805 --> 00:33:38.005
<v Speaker 0>there's a a little bit of a catch.

00:33:38.005 --> 00:33:39.765
<v Speaker 0>There are two billion FPS camera can only

00:33:39.765 --> 00:33:41.950
<v Speaker 0>record videos in one by one pixels. So

00:33:41.950 --> 00:33:43.630
<v Speaker 0>he can so since he can only record

00:33:43.630 --> 00:33:45.230
<v Speaker 0>one pixel at a time, he has to

00:33:45.230 --> 00:33:47.070
<v Speaker 0>keep pointing the camera in different directions, which

00:33:47.070 --> 00:33:49.630
<v Speaker 0>will allow him to tile the image. Tiling

00:33:49.630 --> 00:33:52.610
<v Speaker 0>is something astrophotographers do to get high resolution

00:33:52.669 --> 00:33:55.149
<v Speaker 0>images of celestial objects by combining lots of

00:33:55.149 --> 00:33:59.525
<v Speaker 0>photos they have taken. Wow. So that sounds

00:33:59.525 --> 00:34:03.365
<v Speaker 0>like a lot of work, and it's kind

00:34:03.525 --> 00:34:05.605
<v Speaker 0>some of the scientists, that have seen this

00:34:05.605 --> 00:34:07.285
<v Speaker 0>are arguing that the camera still gives a

00:34:07.285 --> 00:34:09.925
<v Speaker 0>rep still gives a representation of a true

00:34:09.925 --> 00:34:12.665
<v Speaker 0>two billion FPS camera while also having stupendous

00:34:15.030 --> 00:34:15.350
<v Speaker 0>amount of while also saving a stupendous amount

00:34:15.350 --> 00:34:18.870
<v Speaker 0>of money. So that's kinda cool. If you

00:34:18.870 --> 00:34:20.470
<v Speaker 0>wanna check this out, there's a video actually

00:34:20.470 --> 00:34:23.030
<v Speaker 0>linked on the page. So check this page

00:34:23.030 --> 00:34:25.190
<v Speaker 0>out. Go to the article and and and

00:34:25.190 --> 00:34:30.125
<v Speaker 0>watch the video. I, you know, this is

00:34:30.125 --> 00:34:32.845
<v Speaker 0>new. This is this is this is new.

00:34:32.845 --> 00:34:37.565
<v Speaker 0>You know? Next up, I got a space

00:34:37.565 --> 00:34:42.329
<v Speaker 0>article for you folks. Japanese Japan launches advanced

00:34:42.329 --> 00:34:46.170
<v Speaker 0>cargo spacecraft to the ISS. So I know

00:34:46.170 --> 00:34:48.410
<v Speaker 0>that the ISS is is getting decommissioned, but,

00:34:48.410 --> 00:34:51.290
<v Speaker 0>you know, they're still doing missions. Right? Japan's

00:34:51.290 --> 00:34:55.495
<v Speaker 0>aerospace exploration agency, JAXA, has successfully launched its

00:34:55.495 --> 00:34:58.935
<v Speaker 0>new HTVX cargo spacecraft to the International Space

00:34:58.935 --> 00:35:02.295
<v Speaker 0>Station. The mission test upgraded autonomous docking and

00:35:02.295 --> 00:35:05.175
<v Speaker 0>extended payload systems ahead of the ISS retirement

00:35:05.175 --> 00:35:07.515
<v Speaker 0>in twenty thirty. Okay. Yeah. Well, there's that.

00:35:08.480 --> 00:35:10.640
<v Speaker 0>With NASA and other partners winding down ISS

00:35:10.640 --> 00:35:13.119
<v Speaker 0>operations, Japan's commitment signals the start of a

00:35:13.119 --> 00:35:15.520
<v Speaker 0>new commercial era in orbit, one that could

00:35:15.520 --> 00:35:17.920
<v Speaker 0>define how nations share low earth space in

00:35:17.920 --> 00:35:26.224
<v Speaker 0>the next decade. So we'll kinda see how

00:35:26.224 --> 00:35:27.744
<v Speaker 0>it goes. I know that they're planning to

00:35:27.744 --> 00:35:33.525
<v Speaker 0>privatize the, the mission of the ISS to

00:35:33.665 --> 00:35:37.050
<v Speaker 0>companies that, are are basically would be commissioned

00:35:37.050 --> 00:35:45.930
<v Speaker 0>to do the research. So I wonder what

00:35:45.930 --> 00:35:47.610
<v Speaker 0>kind of implications this has. You know? They're

00:35:47.610 --> 00:35:53.445
<v Speaker 0>continuing to con you know, refine their technology.

00:35:54.545 --> 00:35:56.145
<v Speaker 0>I'm sure Japan's gonna have one of their

00:35:56.145 --> 00:35:59.204
<v Speaker 0>own space stations up there doing research. So

00:35:59.265 --> 00:36:01.265
<v Speaker 0>we'll kinda see how that goes, but they

00:36:01.265 --> 00:36:04.625
<v Speaker 0>did launch the this spacecraft on Saturday, October

00:36:04.625 --> 00:36:08.180
<v Speaker 0>twenty fifth. That is today not hot? Did

00:36:08.180 --> 00:36:10.260
<v Speaker 0>I say October twenty fifth earlier? Today's the

00:36:10.260 --> 00:36:17.480
<v Speaker 0>twenty sixth, but so they lifted off at,

00:36:18.420 --> 00:36:20.744
<v Speaker 0>nine AM local time in Japan, but it

00:36:20.744 --> 00:36:23.944
<v Speaker 0>was eight PM eastern time. And it's bex

00:36:24.025 --> 00:36:26.345
<v Speaker 0>expected to arrive at the station for its

00:36:26.345 --> 00:36:29.065
<v Speaker 0>capture and birthing on Wednesday, October twenty ninth

00:36:29.065 --> 00:36:35.860
<v Speaker 0>at about eleven fifty AM EDT. So what

00:36:35.860 --> 00:36:39.800
<v Speaker 0>did they change about this? HTVX enhances transportation

00:36:40.020 --> 00:36:42.580
<v Speaker 0>capabilities and has the capability to provide various

00:36:42.580 --> 00:36:45.540
<v Speaker 0>users with on orbit demonstration opportunities for up

00:36:45.540 --> 00:36:48.260
<v Speaker 0>to one point five years after leaving ISS

00:36:48.260 --> 00:36:52.045
<v Speaker 0>until reentry. Mitsubishi Heavy Industries, which builds the

00:36:52.045 --> 00:36:55.005
<v Speaker 0>h HTB x four JAXA, wrote in the

00:36:55.005 --> 00:37:03.260
<v Speaker 0>description of the vehicle. So it can rough

00:37:03.339 --> 00:37:05.420
<v Speaker 0>it can loft roughly the same payload mass

00:37:05.420 --> 00:37:08.060
<v Speaker 0>to low Earth orbit, about three thousand two

00:37:08.060 --> 00:37:14.160
<v Speaker 0>hundred pounds or six thousand kilograms. And then

00:37:15.725 --> 00:37:19.745
<v Speaker 0>it's gonna the HTVX debut increases the stable

00:37:20.285 --> 00:37:23.085
<v Speaker 0>the stable of our ISS cargo craft by

00:37:23.085 --> 00:37:27.105
<v Speaker 0>one third. The currently operational freighters are Russia's

00:37:27.165 --> 00:37:31.160
<v Speaker 0>Progress vehicle and Cygnus and Dragon spacecraft built

00:37:31.160 --> 00:37:34.700
<v Speaker 0>by American companies Northrop Grumman and SpaceX, respectively.

00:37:35.319 --> 00:37:38.619
<v Speaker 0>Only Dragon is reusable. The others, including HTVX,

00:37:38.839 --> 00:37:40.859
<v Speaker 0>are designed to burn up in Earth's atmosphere

00:37:41.000 --> 00:37:44.119
<v Speaker 0>when their missions are over. Wow. So, you

00:37:44.119 --> 00:37:48.405
<v Speaker 0>know, one person saving a lot of money

00:37:48.405 --> 00:37:52.805
<v Speaker 0>here. Definitely, SpaceX got an advantage with their

00:37:52.805 --> 00:37:58.325
<v Speaker 0>reusable ships. So, cool. That's a pretty sweet

00:37:58.325 --> 00:38:00.660
<v Speaker 0>article. Just some on on some updates on

00:38:00.740 --> 00:38:04.260
<v Speaker 0>what's going on with our space developments. It's

00:38:04.260 --> 00:38:06.819
<v Speaker 0>kinda cool to see continued development of the

00:38:06.819 --> 00:38:11.380
<v Speaker 0>spacecrafts. But I think that's all I got

00:38:11.380 --> 00:38:13.724
<v Speaker 0>for you guys tonight. Thank you guys so

00:38:13.724 --> 00:38:15.984
<v Speaker 0>much for tuning in again for another episode,

00:38:16.525 --> 00:38:19.325
<v Speaker 0>and we're sorry for the the bad scheduling

00:38:19.325 --> 00:38:22.204
<v Speaker 0>this week. We're weren't really clear how we

00:38:22.204 --> 00:38:24.125
<v Speaker 0>were gonna flip back after we did our

00:38:24.125 --> 00:38:26.540
<v Speaker 0>little show swap. I think we're gonna aim

00:38:26.540 --> 00:38:29.340
<v Speaker 0>to kinda get back into that Monday and

00:38:29.340 --> 00:38:34.300
<v Speaker 0>Thursday posting schedule. So so let us know

00:38:34.300 --> 00:38:35.740
<v Speaker 0>what you guys think. If you guys even

00:38:35.740 --> 00:38:38.560
<v Speaker 0>care about the consistency or if you guys

00:38:39.905 --> 00:38:41.905
<v Speaker 0>wanna supposed more shows. You know, we can

00:38:41.905 --> 00:38:44.005
<v Speaker 0>try to do something like that too, but,

00:38:44.785 --> 00:38:46.145
<v Speaker 0>I wanna I wanna know what you guys

00:38:46.145 --> 00:38:48.065
<v Speaker 0>wanna wanna hear. If you guys don't like

00:38:48.065 --> 00:38:50.785
<v Speaker 0>the articles I brought, let me know. I'll

00:38:50.785 --> 00:38:53.025
<v Speaker 0>try to find other other things outside of

00:38:53.025 --> 00:38:54.970
<v Speaker 0>AI. I know we talked about a lot

00:38:54.970 --> 00:39:00.090
<v Speaker 0>of AI stuff today. So if you are

00:39:00.090 --> 00:39:02.830
<v Speaker 0>not a AI listening enthusiast, I am sorry.

00:39:04.010 --> 00:39:05.690
<v Speaker 0>Just a lot of the news today or

00:39:05.690 --> 00:39:08.145
<v Speaker 0>this week has been surrounding that and the

00:39:08.145 --> 00:39:10.785
<v Speaker 0>AWS stuff, of course. But I guess even

00:39:10.785 --> 00:39:13.505
<v Speaker 0>that was kind of built on the the

00:39:13.505 --> 00:39:16.865
<v Speaker 0>AI topic. So, check out the articles. Look

00:39:16.865 --> 00:39:18.865
<v Speaker 0>forward to seeing the newsletter for me. I'm

00:39:18.865 --> 00:39:21.425
<v Speaker 0>gonna be sending that out this, this evening

00:39:21.425 --> 00:39:24.910
<v Speaker 0>probably. Thank you guys again for listening. Have

00:39:24.910 --> 00:39:27.950
<v Speaker 0>a phenomenal night or day or whenever you're

00:39:27.950 --> 00:39:31.230
<v Speaker 0>listening to this. Stay safe out there. Keep

00:39:31.230 --> 00:39:32.990
<v Speaker 0>your heads on straight, and we'll see you

00:39:32.990 --> 00:39:35.644
<v Speaker 0>guys next time. Have an amazing night or

00:39:35.644 --> 00:39:37.804
<v Speaker 0>day. Man, I I don't know what does